For decades, buying stock in hot startups meant being able to invest in funds run by top VCs. But as the AI boom sparks an investment frenzy, more family offices and private assets are bypassing VC intermediaries and entering directly into the cap table.
Mitch Stein, founder of high-net-worth investment advisory firm Arena Private Wealth, told TechCrunch on a recent episode of Equity: “Companies are going private for a long time, and the number of IPOs is historically low.” “A lot of money is being made long before companies go public, and the private market is now dominated by many of these AI companies. Family offices that are allocating money[directly to AI startups]are doing just that.”
Arena recently co-led a $230 million round in AI chip startup Positron, an investment that earned the Midwest company a board seat. Stein says this is part of a deliberate shift from passive allocators to “active participants in the capital markets.”
The urgency for today’s family offices is real.
“The world’s AI infrastructure is being built right now, so you’ll either get in early and have the opportunity to make more primary investments and really build your portfolio, or you’ll miss out on it and make random bets,” Ari Schottenstein, Arena’s head of alternatives, told TechCrunch.
Mr. Stein put it more bluntly: “The biggest risk is not what happens to your investment in AI, but rather that you don’t touch it.”
tech crunch event
San Francisco, California
|
October 13-15, 2026
The numbers reflect this sentiment. In February, family offices made 41 direct investments in startups, almost all of them related to AI. These include notable companies like Laurene Powell Jobs’ Emerson Collective to World Labs, Azim Premji’s Family Office to Runway, and Eric Schmidt’s Hillspire to Goodfire. According to a study by BNY Wealth, 83% of family offices say AI is a top strategic priority over the next five years, and more than half have exposure to AI through investments.
Some go even further. Schottenstein said many family offices are incubating their own AI companies, seeding the first million companies, taking on operational roles and deploying the same entrepreneurial instincts that first made them rich. Jeff Bezos’ decision to become CEO of his robotics company, which initially raised $6.2 billion last year at a valuation of nearly $30 billion, is a high-profile example of this model.
On a smaller scale, Stein cited Tyson Tuttle, an Austin-based angel investor and former CEO of Silicon Labs, which Texas Instruments agreed to acquire for $7.5 billion. Tuttle co-founded Circuit, a startup that uses AI to improve manufacturing and distribution, and raised a $30 million angel round, including $5 million from his family office.
However, not everyone who comes to the table founded a company. Arena’s team comes from an institutional finance background and claims that earning the right to lead a round is through rigorous due diligence.
“We take our time and are very slow to say ‘yes’ and often say ‘no,'” Schottenstein said. “We will definitely invest in the resources, experts and talent necessary to make sure that companies are what they say they are and can do what they say they will do.”
In the case of the Positron deal, this means working with third-party experts to validate the technology, but it also means reading the cap table itself as a signal. “If Arm is going to sign a contract, we’d like to think your technology is real,” Schottenstein said. Arena also knows that Oracle is a major customer, making Positron one of the only AI chips to go into a hyperscaler not named Nvidia or AMD.
This selectivity determines how arenas participate once they have joined. Unlike a typical VC, which spreads risk across a portfolio, Arena makes a small number of direct deals per year, completely changing the stakes. When they’re on board, everyone’s on board. Positron is the company’s only AI inference chip investment.
“If we participate in single-asset direct trading and only make a small number of trades each year, our stakes are incredibly high,” Stein said. “We don’t manage returns at the portfolio level. We don’t model failure on single asset trades. We take on a lot of risk with concentrated customer capital. We take on reputational risk as a company. We allocate a tremendous amount of time and resources. There’s an integrity there that our founders appreciate.”
